How fast does Gemma 4 E4B run on the Apple M1 Ultra?

Gemma 4 E4B runs at about 69.7 tokens/s on the Apple M1 Ultra (measured, Q4_K_M, Metal, llama-bench). Based on 1 measured run(s), each linked to its public source.

Gemma 4 E4B › Apple M1 Ultra

Gemma 4 E4B on the Apple M1 Ultra

Gemma 4 E4B runs at about 69.7 tokens/s on the Apple M1 Ultra (measured, Q4_K_M, Metal, llama-bench).

Best measured: 69.7 tokens/s generation and 1,173 tokens/s prompt processing (Q4_K_M, Metal).

Every run

Quant	Backend	Generate tok/s	Prompt tok/s	Test	Build / by / notes	Source
Q4_K_M	Metal	69.7	1,173	tg128	ff5ef8278 Fazal Majid llama-bench --hf-repo unsloth/gemma-4-E4B-it-GGUF; GPU core count not stated	Fazal Majid llama-bench log (utumno)

tg128 = generating 128 tokens; pp512 = reading a 512-token prompt (llama-bench). LocalScore averages several prompt and output lengths, so its numbers run lower than tg128. "decode, llama-server" = single-stream generation speed measured against a running llama-server (no speculative decoding).

Check prices

Amazon

As an Amazon Associate I earn from qualifying purchases.

No card yet?

Try it on a rented GPU (RunPod)

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.

All results for the Apple M1 Ultra · Gemma 4 E4B on other hardware

Embed this result

HTML

<a href="https://rigbench.pages.dev/model/gemma-4-e4b-it/on/apple-m1-ultra/"><img src="https://rigbench.pages.dev/badge/gemma-4-e4b-it--apple-m1-ultra.svg" alt="Gemma 4 E4B · Apple M1 Ultra: 69.7 tok/s (Q4_K_M, measured)"></a>

Markdown (GitHub README, forums)

[![Gemma 4 E4B · Apple M1 Ultra: 69.7 tok/s (Q4_K_M, measured)](https://rigbench.pages.dev/badge/gemma-4-e4b-it--apple-m1-ultra.svg)](https://rigbench.pages.dev/model/gemma-4-e4b-it/on/apple-m1-ultra/)